Last released Oct 14, 2025
Pythonic LLM inference on legacy GPUs using Vulkan — GPU-accelerated local AI for AMD, Intel, and NVIDIA without CUDA.
Supported by